Estrogen receptor α (ERα) and estrogen receptor β (ERβ) are ligand-regulated nuclear receptor transcription factors. Both receptors control important biological processes in humans and other eukaryotes through regulating the expression of genes involved in development, reproduction, bone homeostasis, brain function, metabolism, and inflammation (Heldring et al., 2007; Jia et al., 2015; Lee et al., 2012). The two receptors are differentially expressed between tissues. ERα is expressed in female reproductive tissues, breast, bone, lung, kidney, liver, and brain, while ERβ is expressed in both male and female reproductive tissues, spleen, bone, lung, kidney, colon, brain, and the immune system (Bord et al., 2001; Brandenberger et al., 1997; Couse et al., 1997; Jia et al., 2015). This supports that the receptors play distinct, tissue-specific roles that vary dependent on the stage of development and changing physiological conditions. The same overall molecular architecture is preserved for ERα and ERβ, comprised of an N-terminal domain (NTD), a central DNA-binding domain (DBD) and a C-terminal ligand binding domain (LBD). The NTD contains the activation function 1 (AF1) region involved in coregulator recognition and functions as a ligand-independent transactivation, but this region is poorly conserved between ERα and ERβ with only 25 % sequence identity. The DBD is highly conserved between each of the receptors and acts to recognize specific DNA sequences termed estrogen response elements (EREs) in the promoter region of target gene. Finally, the LBD confers ligand-dependent regulation of transcription through controlling formation of the activation function 2 (AF2) surface, which allows recruitment of coregulators to the receptor, commonly via an LxxLL motif. The ERα and ERβ LBDs share moderate sequence identity (59 %) but have highly conserved structures. Despite the high level of structural conservation, some subtype selective ligands have been identified (Paterni et al., 2014). Under physiological conditions the activity of ERα and ERβ is directly regulated by the primary endogenous estrogens which behave as agonists of the receptor, the most abundant being estradiol (E2). The serum concentration of E2 is typically in the picomolar range, being lower in men and post-menopausal women (∼1–150 pM) and elevated in pre-menopausal women (50–1500 pM) (Frederiksen et al., 2019; Mayo Clinic Laboratories, 2025). The classical mechanism of ER activation begins with E2 binding to the receptor LBD, which exists in an inactive state complex with chaperones prior to ligand binding (Diel, 2002). This may occur within the nucleus or cytoplasm, and results in translocation and accumulation of the receptor within the nucleus (Kocanova et al., 2010; Moriyama et al., 2020). Ligand binding leads to homodimerization of the receptors via the LDB and DBD, which then bind to EREs contained in target gene promoters. Binding of E2 also results in a structural rearrangement of the receptor ligand binding domain which facilitates the formation of the AF2 and recruitment of coactivators, activating transcription.

Aside from the primary estrogens, other endogenous steroid metabolites have been identified which also modulate the activity of ERs. Androst-5-ene-3β,17β-diol, herein referred to as androstenediol, is one such example, being an intermediate in the biosynthesis of testosterone from dehydroepiandrosterone (DHEA) (Fig. 1). Despite sharing a closer chemical structure to the primary androgens testosterone and dihydrotestosterone (DHT), it has been determined that androstenediol has more potent estrogenic activity than androgenic activity. This estrogenic activity was recognized due to the ability of androstenediol to produce estrogenic responses in human MCF-7 breast cancer cells in an ER-dependent manner (Adams et al., 1981; Adams, 1985; Boccuzzi et al., 1994). While less potent than E2, these estrogenic responses were achieved at low nanomolar concentrations of androstenediol. As the serum and plasma concentration of androstenediol are reported to be in the low nanomolar range (1–5 nM) in normal men and women, this supports that the estrogenic activity of androstenediol is relevant in a physiological setting (Rosenfield and Otto, 1972; Tagawa et al., 2001). Transcriptional reporter assays revealed that androstenediol behaves as an agonist toward both ERα and ERβ (Michael Miller et al., 2013). In vitro competitive binding assays have also revealed that the estrogenic activity of androstenediol is elicited through direct binding to the ER LBD (Kuiper et al., 1997; Nicoletti et al., 2010; Chen et al., 2013). Kuiper et al. reported that, while E2 binds to ERα and ERβ with a similar potent affinity (Ki ∼ 100 pM), androstenediol displays a weaker interaction, with relative affinities compared to E2 of 6 % and 17 % for the respective receptors (Kuiper et al., 1997). Intriguingly, this also revealed that androstenediol possesses a ∼3-fold selectivity for ERβ over ERα, a property not displayed by the primary endogenous estrogens, including E2.

Since this discovery, it has been identified that the selectivity of androstenediol for ERβ may be highly relevant in certain physiological contexts. Saijo et al. reported that androstenediol can act as a modulator of ERβ to supress proinflammatory responses in vitro in lipopolysaccharide (LPS) stimulated microglia and astrocytes, and in vivo in a mouse model of multiple sclerosis, raising the possibility that androstenediol may act in vivo to regulate inflammation within the central nervous system (Saijo et al., 2011). This was discovered to occur by a transrepression mechanism that does not follow the classical mechanism of ER activation. Instead, androstenediol was found to tether ERβ to the activator protein-1 (AP-1) transcription factor cFos, which facilitates recruitment of C-terminal binding protein (CtBP) corepressor complexes to AP-1 dependent promoters. While a subset of synthetic ERβ-selective ligands produced a similar effect in repressing proinflammatory responses, E2 did not. Given that E2 and androstenediol are both agonists of ERα and ERβ, it remains unclear as to why these differential responses were observed. A possible explanation for this discrepancy is that these hormones exert different effects on the structure of ERβ. However, no crystal structures of either ERα or ERβ bound to androstenediol have been reported to date. In this work we used X-ray crystallography to investigate whether differential binding modes of E2 and androstenediol within human ERα and ERβ (hERα and hERβ) LBDs might be responsible for this unique behaviour by i) identifying the binding modes of androstenediol with hERα and hERβ LBDs and ii) determining the basis of selective binding for androstenediol to the hERβ LBD.